Abstract :

Chest X-ray Imaging Remains One Of The Most Widely Used And Cost-effective Diagnostic Tools For Respiratory Diseases, Yet Manual Interpretation Is Often Challenging Due To Overlapping Anatomical Structures, Interobserver Variability, And The Subtle Nature Of Certain Abnormalities. Recent Advancements In Deep Learning Have Enabled Highly Accurate Automated Detection Systems Capable Of Supporting Radiologists In Clinical Decisionmaking. This Research Proposes A Novel Deep Learning–enhanced Framework That Integrates Multi-scale Convolutional Feature Extraction, Attention-driven Localization, And Domainadapted Training To Improve Disease Classification Performance Across Diverse Chest X-ray Datasets. The Proposed System Aims To Address Limitations Observed In Traditional Single-scale CNN Architectures By Incorporating Hierarchical Learning Layers That Capture Both Fine-grained Pathological Markers And Global Thoracic Patterns. Furthermore, An Adaptive Data Augmentation Pipeline And Balanced Loss Optimization Are Implemented To Mitigate Class Imbalance And Reduce Overfitting, Ensuring Improved Robustness Across Real-world Imaging Conditions. Experimental Evaluations Demonstrate That The Framework Achieves Superior Accuracy, Sensitivity, And Specificity Compared To Baseline Models. The Study Highlights The Potential Of Deep Learning As A Transformative Tool For Scalable Screening Of Diseases Such As Pneumonia, Tuberculosis, Lung Opacity, Cardiomegaly, And COVID-19. This Research Contributes To The Growing Field Of Medical Image Analysis By Presenting An Enhanced Architecture Capable Of Supporting Radiological Workflows And Improving Diagnostic Outcomes, Especially In Resource-limited Healthcare Settings. Keywords: Deep Learning, Chest X-Ray Analysis, Disease Detection, Convolutional Neural Networks, Medical Image Processing, Attention Mechanisms, Automated Diagnosis

Published:

12-12-2025

Issue:

Vol. 25 No. 12 (2025)

Page Nos:

207-215

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite